Text Generation
Transformers
Safetensors
busybeaver_qdelta
busybeaver
tool-calling
agent-policy
json
local-agents
qdelta
50m
Instructions to use GestaltLabs/BusyBeaver-50M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use GestaltLabs/BusyBeaver-50M with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="GestaltLabs/BusyBeaver-50M")# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("GestaltLabs/BusyBeaver-50M", dtype="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- vLLM
How to use GestaltLabs/BusyBeaver-50M with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "GestaltLabs/BusyBeaver-50M" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "GestaltLabs/BusyBeaver-50M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/GestaltLabs/BusyBeaver-50M
- SGLang
How to use GestaltLabs/BusyBeaver-50M with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "GestaltLabs/BusyBeaver-50M" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "GestaltLabs/BusyBeaver-50M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "GestaltLabs/BusyBeaver-50M" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "GestaltLabs/BusyBeaver-50M", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use GestaltLabs/BusyBeaver-50M with Docker Model Runner:
docker model run hf.co/GestaltLabs/BusyBeaver-50M
Upload V12 resolved eval artifacts
Browse files
eval/evaluation_report_v12_path_grounding_resolved_full.md
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# BusyBeaver Evaluation Report
|
| 2 |
+
|
| 3 |
+
- Mode: model
|
| 4 |
+
- Rows: 384
|
| 5 |
+
|
| 6 |
+
- json_validity_rate: 1.0000
|
| 7 |
+
- strict_json_rate: 1.0000
|
| 8 |
+
- schema_validity_rate: 1.0000
|
| 9 |
+
- valid_tool_rate: 1.0000
|
| 10 |
+
- correct_tool_accuracy: 1.0000
|
| 11 |
+
- argument_exact_match: 0.8958
|
| 12 |
+
- argument_semantic_match: 0.9792
|
| 13 |
+
- unnecessary_escalation_rate: 0.0000
|
| 14 |
+
- unsafe_command_rate: 0.0000
|
| 15 |
+
- placeholder_rate: 0.0000
|
| 16 |
+
- correct_tool_and_arg_semantic: 0.9792
|
| 17 |
+
- repeated_action_loop_rate: 0.0000
|
| 18 |
+
- concrete_argument_semantic_match: 0.9792
|
| 19 |
+
- recovery_action_accuracy: 1.0000
|
| 20 |
+
- concrete_argument_rows: 384.0000
|
| 21 |
+
- recovery_rows: 128.0000
|
| 22 |
+
|
| 23 |
+
## Grouped Metrics
|
| 24 |
+
|
| 25 |
+
### edit (n=32)
|
| 26 |
+
- json_validity_rate: 1.0000
|
| 27 |
+
- strict_json_rate: 1.0000
|
| 28 |
+
- schema_validity_rate: 1.0000
|
| 29 |
+
- valid_tool_rate: 1.0000
|
| 30 |
+
- correct_tool_accuracy: 1.0000
|
| 31 |
+
- argument_exact_match: 1.0000
|
| 32 |
+
- argument_semantic_match: 1.0000
|
| 33 |
+
### escalate (n=32)
|
| 34 |
+
- json_validity_rate: 1.0000
|
| 35 |
+
- strict_json_rate: 1.0000
|
| 36 |
+
- schema_validity_rate: 1.0000
|
| 37 |
+
- valid_tool_rate: 1.0000
|
| 38 |
+
- correct_tool_accuracy: 1.0000
|
| 39 |
+
- argument_exact_match: 1.0000
|
| 40 |
+
- argument_semantic_match: 1.0000
|
| 41 |
+
### execute (n=32)
|
| 42 |
+
- json_validity_rate: 1.0000
|
| 43 |
+
- strict_json_rate: 1.0000
|
| 44 |
+
- schema_validity_rate: 1.0000
|
| 45 |
+
- valid_tool_rate: 1.0000
|
| 46 |
+
- correct_tool_accuracy: 1.0000
|
| 47 |
+
- argument_exact_match: 1.0000
|
| 48 |
+
- argument_semantic_match: 1.0000
|
| 49 |
+
### inspect (n=128)
|
| 50 |
+
- json_validity_rate: 1.0000
|
| 51 |
+
- strict_json_rate: 1.0000
|
| 52 |
+
- schema_validity_rate: 1.0000
|
| 53 |
+
- valid_tool_rate: 1.0000
|
| 54 |
+
- correct_tool_accuracy: 1.0000
|
| 55 |
+
- argument_exact_match: 0.9922
|
| 56 |
+
- argument_semantic_match: 0.9922
|
| 57 |
+
### memory (n=32)
|
| 58 |
+
- json_validity_rate: 1.0000
|
| 59 |
+
- strict_json_rate: 1.0000
|
| 60 |
+
- schema_validity_rate: 1.0000
|
| 61 |
+
- valid_tool_rate: 1.0000
|
| 62 |
+
- correct_tool_accuracy: 1.0000
|
| 63 |
+
- argument_exact_match: 0.7812
|
| 64 |
+
- argument_semantic_match: 0.7812
|
| 65 |
+
### other (n=96)
|
| 66 |
+
- json_validity_rate: 1.0000
|
| 67 |
+
- strict_json_rate: 1.0000
|
| 68 |
+
- schema_validity_rate: 1.0000
|
| 69 |
+
- valid_tool_rate: 1.0000
|
| 70 |
+
- correct_tool_accuracy: 1.0000
|
| 71 |
+
- argument_exact_match: 0.6667
|
| 72 |
+
- argument_semantic_match: 1.0000
|
| 73 |
+
### test (n=32)
|
| 74 |
+
- json_validity_rate: 1.0000
|
| 75 |
+
- strict_json_rate: 1.0000
|
| 76 |
+
- schema_validity_rate: 1.0000
|
| 77 |
+
- valid_tool_rate: 1.0000
|
| 78 |
+
- correct_tool_accuracy: 1.0000
|
| 79 |
+
- argument_exact_match: 1.0000
|
| 80 |
+
- argument_semantic_match: 1.0000
|